91

programming is terrible

lessons learned from a life wasted

@tef

hello, iamtef, I am a bad programmerI took things apart to find out how they work. Now I am surprised they do.

I am a bad programmer (tests, docs, apology code review, bugs)

we can do better.

2

Good and Bad ProgrammersHow Culture Dictates CodeIndoctrination vs Learning

Being Successful vs Being Good

Not just complaints, Tips, hope.

mistakes i’ve made etc. mythos.

people rather than code.

std disclaimer, opinions, not work, or reality

3

YMMVHTH

HAND

Some find useful, gives me trouble.

I am wrong. Doesn’t stop bloggers.

A good place to start

4

my code is better than your code

(sing it)

false dichotomy of good and bad programmers.

they normally mean this

5

Programmers who are like me.

Programmers who are not like me.

cargo cult their personality -> success6

Programmers who use my favourite language.

Programmers who don’t.

blub paradox.

aside: unmaintainable lisp, was replaced with perl.aside: lisp would have prevented 9/11

imply good or bad.

7

Programmers who share my political views.

Programmers who don’t.

mash types into politics for people who understand neither.

why? it is easy. simple answer to a hard question.also blog hits. especially emotionally charged/trolling

8

10 times50 times100 times

the myth of genius. rockstar, ninja, founder, entrepreneurdrowning in puberty/machismoall bollocks, faulty study, repeated ad-nauseumnothing to learn cos they’re smart or thick‘A type programmer’ = ‘Easily exploited hard worker’Worst good/bad

9

Programmers who are men.

Programmers who aren’t.

kill yourself now.

Trials on two groups about belief in magic penis abilities.You are bad person and bad programmingtabs vs spaces joke.

so, if there are two types, this is what I think it looks like

10

Programmers who know they make mistakes.

Programmers who think they don’t

A little of A, a little of B. Sometimes refuse to try, sometimes refuse to learn.Biggest mistake is optimism.

11

“You would think that...”

Optimisim is still necessary or we would go mad.Chronically underestimate work.classic programmer compliaint,underling optimism,could be better, or they could do better.I am a cynic. We fix bugs but not people.

12

PEBKAC

Mistakes come from environment too.We need to find out why bug happenedIf we want to stop writing so many bugs

Not endemic, systematic.

13

“...organizations which design systems ... are constrained to produce designs which are copies of the communication structures of these organizations”

Code reflects social structures.You want a service, the team needs to provide/run it,not hand over code.God objects come from god programmers.Groups make mistakes too.

14

The Bike Shed

design of nuclear power plantparkinson’s law.when domain expertise isn’t needed, everyone is an experteveryone has an opinion, debate will never endputting your thumb on it, etc

15

The Group Project

let’s do stuff together, pool out ideas.ideas no-one wants to actually d on their owncolaboration requires leading by exampledoesn’t mean ideas, ideas are multiplierif ideas are all you have

16

The ‘Goon’ Project

We have a wiki, and a logo. All bikeshedding, no progress.When you have enthusiam and not much else.Let’s write a video game.Let’s build a darknet.Sure, i’ll make the wiki and the logo.

17

Waterfall

Not just people or teams, but methodologies.Waterfall was introduced as a strawman by royce.People read the first page, eh. that’ll do

Project management is often control over measurement or feedbackMilestones handed out, we’ll just have to do it perfectly.

18

Manhole Cover

Also how we get workers too. We can’t find good programmers.We can’t interview, or measure it. We end up doing brainteasers.What should you do when you get a brainteaser?

19

LEAVE

Wasson Selection TaskNot a quiz show host. See how you cope with stupid questionsIt’s good because our management are terrible.But FACEBOOK,GOOGLE, MICROSOFTAny company that does this is universally bad.

20

Occultism

Not science, Art, but rituals, cargo culting. Best practices are superstition. Say we’re scientists, but don’t test code, methologies, or assumptions.At best, a craft.We learn more from maintenance, fix, test, adapt, evolving. Than creating.Not to deny experiments, ‘top down the second time’ still true. Prototypes help you explore an idea, maintanance brings understanding.Not just bad code, bad practices, but bad teaching

21

Teaching

Nostalgia and learning preferences. Do what I did and learn what I did.“What must they know” rather than “What do they want to learn and how?”Learning Preferences matter. Need to encourage people to learn and explore on own , rather than dictating course. Guidance and support is good. Adult vs child education. If asked what to learn, I ask what they want to create.

22

my first language

they can do something fun with in an afteroon,friends know and support, easy to install, doesn’trequire learning ides. scripting languages.

don’t worry about OO, or advanced things. getsomething simple running and play.learning programming should be a step of something fun

23

Learning Through Play

Find a sandbox, turtle graphics, music. Something where you get instant feedback,and easy tweaking. Get them to explain things to me, and questioningthan rote/memorization (PUB STAT VOID MAIN). Computer is a tool for exploring ideas, not a beautracraft with form filling.

Guilty: I am nostalgic too. For Logo

24

Seymour PapertMindstorms

LOGO, Math World, Users build rules, no rules but their own.Lets them see consequences, understand things (verb stuff)

Similar idea in math. death march through formulae, not actual problemsolving. Learning is fun when you get to be creative.

25

view-source

other influence, get to play, get to change, get to tweakprogramming is not just a way to do business rules,but a way to understand things.

some bloggers (terrible atwood) tells people off for playing and learningas they did, and they are bad people.

26

REAL MEN USE C

C is useful, uses and popular. It’s also hard,demanding. Not a good first lang.

Hard to do simple things quickly. C is character buildingShell before C.

27

USED IN INDUSTRY

C#/Java always simpler script instead atop runtime.OO is hard to grok without learning why it is useful.Hard to begin without forced ignorance.Berter second, not first.Aside: Learning with caution, looking for excuse to avoid,like on forums, want to avoid effort.

28

MATHS IS HARD

Maths/{prpgramming are related. Also floating pointmisconceptsions. Need enough maths as your problem demands.Not many programs needs things beyond counting. Spreadsheet knowhow good enough.Programming is part maths, not diff-geom, but same dicipline of thought, reasoning,Programmers don’t need to be maths, but are maths. Curry Howard, innit.Ultimatelt interdiciplinary, writing, critical reason, engineer dicipline, maths reasoning.Overlooked, domain experience of problem.

29

How to be successful

I’m not good enough. I go mad. Other people seem to survive,and I will share strategies. Document badly, ignoring failure cases, types or args. too busy to share info, be writing code. Lots of code. With autocomplete. Wrappers, avoid stdlib.Revinent, go crazy with adv. feat. Modules, abstracton everywhere, esp if it only makes sense when combined.Fix bugs by making new ones. Close bugs that get reopened byt someone else. Shifting problem. Non-det tests, hard setup. Be the central point of fail, or be replaceable. Create work for yourself, be a soliphist, sabotage is rewarded.

30

How to be good

Read code, learn from others. Professional writer. Ten years experience. All courses are about making new software not fixing existing things. no focus on maintenance, debuging or analysis.

Estm what hasn’t been done before may be due to ignorance. Write cose as if it is wrong and you will have to delete it. Probably will. Fail quickly, get is wrong. Easier when you don’t write too much code. Empathy good. Don’t be a diva, an artist, don’t try to be right, but make it less painful to be wrong. Easy to replace not easy to extend.

31

A final warning

the software industry is terrible, so is every other industry. retraining won’t help you escape people.

32